search

frankmaayn / galaxy_computer_vision_with_EFIGI_data

0 stars 0 forks source link

Image Preprocessing #7

Open dpalencia opened 3 years ago

dpalencia commented 3 years ago

This can be written as a class as well, will contain a method which takes as input raw image data and preprocesses images by thresholding and extracting keypoints with a SIFT algorithm. This can serve as the input for our GBM model or any SKLearn model we wish to use.

dpalencia commented 3 years ago

We can run the image preprocessing separately and save the results as tabulated data (as a .csv file)

dpalencia commented 3 years ago

Added a script that processes images and uses SIFT to extract features. I used this script to produce .csv file with extracted features from the SIFT algorithm.

This code has room for improvement. Mainly, the code does not work when SIFT identifies fewer keypoints than the number specified in the code. For this reason, I only used 2 keypoints. We can experiment with adding more, and handling missing null values.

In the meantime, we can use these features as inputs for our SIFT pipeline.